CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language...

36
CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE Dept., IIT Bombay 5 th April, 2012 (main text: Syntax by Adrew Carnie, Blackwell Publication, 2002)

Transcript of CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language...

Page 1: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

CS460/626 : Natural Language Processing/Speech, NLP and the Web

(Lecture 35– X-bar theory)

Pushpak BhattacharyyaCSE Dept., IIT Bombay

5th April, 2012

(main text: Syntax by Adrew Carnie, Blackwell Publication, 2002)

Page 2: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

How deep should a tree be?

� Is there a principle in branching

� When should the constituent give rise to children?to children?

� What is the hierarchy building principle?

Page 3: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Deeper trees needed for capturing sentence structure

NP

PPAPThe

This wont do! Flat structure!

PP

big of poems

with the blue cover

[The big book of poems with theBlue cover] is on the table.

book

Page 4: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Other languages

NP

PPAPThe

English

PP

bigof poems

with the blue cover

[niil jilda vaalii kavita kii kitaab]

book

NP

PPAP

niil jilda vaalii kavita kii

kitaab

badii

Hindi

PP

Page 5: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Other languages: contd

NP

PPAPThe

English

PP

bigof poems

with the blue cover

[niil malaat deovaa kavitar bai ti]

book

NP

PPAP

niil malaat deovaa kavitar

bai

motaa

Bengali

PPti

Page 6: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

PPs are at the same level: flat with respect to the head word “book”

NP

PPAPThe

No distinction in terms of dominance or c-command

PP

big of poems

with the blue cover

[The big book of poems with theBlue cover] is on the table.

book

Page 7: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

“Constituency test of Replacement” runs into problems

� One-replacement:

� I bought the big [book of poems with the blue cover] not the small [one]blue cover] not the small [one]

� One-replacement targets book of poems with the blue cover

� Another one-replacement:

� I bought the big [book of poems] with the blue cover not the small [one] with the red cover

� One-replacement targets book of poems

Page 8: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

More deeply embedded structureNP

APThe

N’1

N’2

PP

big

of poems

with the blue coverNbook

PPN’3

Page 9: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

To target N1’

� I want [NPthis [N’big book of poems with the red cover] and not [Nthat [None]]

Page 10: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Bar-level projections� Add intermediate structures

� NP� (D) N’

� N’� (AP) N’ | N’ (PP) | N (PP)

() indicates optionality� () indicates optionality

Page 11: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

New rules produce this treeNP

APThe

N’1

N’2

N-bar

PP

big

of poems

with the blue coverNbook

PPN’3

Page 12: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

As opposed to this tree

NP

PPAPThe

PP

big of poems

with the blue coverbook

Page 13: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

V-bar

� What is the element in verbs corresponding to one-replacement for nounsnouns

� do-so or did-so

Page 14: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

As opposed to this tree

NP

PPAPThe

PP

big of poems

with the blue coverbook

Page 15: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

I [eat beans with a fork]

VP

eat

PP

NP

beans

with a fork

No constituent that groups together V and NP and excludesPP

Page 16: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Need for intermediate constituents

� I [eat beans] with a fork but Ram [does so] with a spoon

VP

V1’

V2’

NP

beans

eat

with a fork

PP

V

VP�V’V’� V’ (PP)V’� V (NP)

Page 17: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

How to target V1’

� I [eat beans with a fork], and Ram [does so] too.

VP

V1’

V2’

NP

beans

eat

with a fork

PP

V

VP�V’V’� V’ (PP)V’� V (NP)

Page 18: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Case of conjunction

V3’PP

VP

V1’

V ’

V2’

V3’

NP

beans

eat

In the afternoon

V

V4’

NP

coffee

drink

V

Conjand

Page 19: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

A-bar: adjectives

AP

A1’

AP

A2’

AP� A’A’� (AP) A’A’� A (PP)

A3’

A4’

blue

Very

bright

A5’

A6’

green

Conjand

AP AP

dull

Page 20: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

So-replacement for adjectives

� Ram is very serious about studies , but less so than Shyam

Page 21: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

P-bar: prepositions

AP

A1’

AP

P1’PP� P’P’� P’ (PP)P’� P (NP)

PP

P2’

NP

the table

right

off

P3’

NP

the trash

Conjand

P P

into

Page 22: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

So-replacement for Prepositions

� Ram is utterly in debt, but Shyam is only partly so.

Page 23: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Complements and Adjuncts orArguments and Adjuncts

Page 24: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Rules in bar notation: Noun

� NP� (D) N’

� N’� (AP) N’

� N’� N’ (PP)� N’� N’ (PP)

� N’� N (PP)

Page 25: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Rules in bar notation: Verb

� VP� V’

� V’� V’ (PP)

� V’� V (NP)� V’� V (NP)

Page 26: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Rules in bar notation: Adjective

� AP� A’

� A’� (AP) A’

� A’� A (PP)� A’� A (PP)

Page 27: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Rules in bar notation: Preposition

� PP� P’

� P’� P’ (PP)

� P’� P (NP)� P’� P (NP)

Page 28: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Introducing the “X factor”

� Let X stand for any category N, V, A, P

� Let XP stand for NP, VP, AP and PP

� Let X’ stand for N’, V’, A’ and P’� Let X’ stand for N’, V’, A’ and P’

Page 29: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

XP to X’

� Collect the first level rules

� NP� (D) N’

� VP� V’� VP� V’

� AP� A’

� PP� P’

� And produce

� XP� (YP) X’

Page 30: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

X’ to X’

� Collect the 2nd level rules

� N’� (AP) N’ or N’ (PP)

� V’� V’ (PP)� V’� V’ (PP)

� A’� (AP) A’

� P’� P’ (PP)

� And produce

� X’� (ZP) X’ or X (ZP)

Page 31: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

X’ to X

� Collect the 3rd level rules

� N’� N (PP)

� V’� V (NP)� V’� V (NP)

� A’� A (PP)

� P’� P (NP)

� And produce

� X’� X (WP)

Page 32: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Basic observations about X and X’

� X’� X (WP)

� X’� X’ (ZP)

� X is called Head� X is called Head

� Phrases must have Heads:Headedness property

� Category of XP and X must match: Endocentricity

Page 33: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Basic observations about X and X’

� X’� X (WP)

� X’� X’ (ZP)

� Sisters of X are complements� Sisters of X are complements

� Roughly correspond to objects

� Sisters of X’ are Adjuncts

� PPs and Adjectives are typical adjuncts

� We have adjunct rules and complement rules

Page 34: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Structural difference between complements and adjuncts

X’

ZP

XP

X’

WP

Complement

X

Adjunct

Page 35: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Complements and Adjuncts in NPs

N’

ZP

NP

N’

PP

of poems

N

with red cover

book

Page 36: CS460/626 : Natural Language Processing/Speech, NLP and ... · CS460/626 : Natural Language Processing/Speech, NLP and the Web (Lecture 35– X-bar theory) Pushpak Bhattacharyya CSE

Any number of AdjunctsN’

ZP

N’

NP

from Oxford Press

N’

PP

of poems

N

with red cover

book